Compendium Federal Technology, LLC — Department of Defense SBIR Phase I: N204-A01

Description

This research will provide the ability to autonomously scan large platforms (e.g., ships, aircraft, and ground vehicles) with great precision. These scans can locate structural anomalies and support creation of digital twins. This proposal addresses the use of Artificial Intelligence (AI) with Deep Neural Networks (DNNs) hosted on portable devices. Commercial-off-the-shelf (COTS) products can be integrated to acquire data, and deploy applications for detection, classification and inferencing, which inspect platforms using real-time video and still images. The miniaturization of microcomputers and Graphics Processing Units (GPUs) has enabled the deployment of low-cost systems that collect data at very high rates and host intelligent applications which detect and classify objects. Infrastructure inspection will be the focus of federal, state and local government investment in the coming decade. The development of state-of-the-art GPUs is being accelerated by automation of manufacturing and logistics systems using robotics as well as the introduction of self-driving cars. These systems require the fusion of multiple sensor streams which collect data and make decisions in real-time. There are several emerging technologies that can be leveraged to yield increases in productivity and achieve significant reductions in maintenance hours by automating inspection procedures using real-time AI combined with smart inspection devices.